Hi all 
,
New Trx !
73, Éric F5JKK
12 Likes
Watch the YouTube video by K4SWL last night, it looks an awesome little rig. Perfect to leave in the car for quick, unplanned activations.
I think the Pebble will be popular. Now we just need Neil @G7UFO to organise a European assembly operation! I could easily assemble a PCB by hand, with a board order from JLC, but I don’t have experience with 3D printing.
The project is open source, with the BOM and gerbers online, but I do think something should be kicked back to the project originators too.
What a great idea for a radio!
73, Colin
9 Likes
Yeah, it does look smart. Have been chatting to Mike, N4FFF about the mic wiring and element for the Pebble HF.
My understanding is that they’re wanting to get a few out of the door and the design nailed down before looking to expand where it’s available.
3 Likes
Saw the videos last night. I’m going to order one when they become available. I have a K6ARK 49:1 unun that is not being used (prefer the 9:1 version with my KX2). This radio, the 49:1 EFHW antenna, and a spare battery from my KX2 will make for a tidy little radio for lightweight activations.
Chris
If you are interested in this project, visit groups.io here 
https://groups.io/g/BluQRP/topic/118455508#msg353
73, Éric F5JKK
I decided to ask Google AI for help.
My Question: Do you think that the creators of the HF pebble should credit Hans Summers for his work on the QCX mini?
Google AI’s Answer:
Whether the creators of the HF Pebble (and the broader uSDX community) owe a formal “IP credit” to Hans Summers (G0UPL) is a nuanced debate within the ham radio community.
From a technical and ethical standpoint, here is how the credit typically breaks down:
1. The Argument for Direct Credit
The Class-E PA and the 74ACT00 bus-switch receiver implementation in the Pebble are virtually identical to the circuit blocks Hans popularized in the QCX.
The “QCX-SSB” Link: The uSDX firmware (which the Pebble runs) was originally written specifically for the QCX hardware.
Engineering Effort: Hans spent years optimizing the specific component values (like the BS170 bias and the feedback loops) to make a $50 radio perform like a $500 one. Using those exact values in a new PCB design is, at the very least, an inherited engineering feat.
2. The Argument for “Common Domain” Building Blocks
From a strict “Intellectual Property” perspective, Hans did not invent the Tayloe Mixer (Dan Tayloe, N7VE, did) or the Class-E Amplifier (the Sokals did in the 70s).
Open Architecture: The Si5351A + ATmega328 + Tayloe Mixer is now considered a “standard architecture” for budget SDRs.
Independent PCB: Because the Pebble is a ground-up PCB layout designed for a different purpose (SDR/SSB vs. Hans’s CW-only focus), many builders view it as an independent project that happens to use the same “industry standard” parts for the QRP price point.
3. The Ethical “Middle Ground”
Most seasoned builders feel that a “Derived From” or “Inspired By” credit is the right thing to do.
The uSDX Precedent: The main uSDX developers generally acknowledge that their software journey began on the QCX.
Community Friction: There has been historical friction because “clones” (unauthorized 1:1 copies of the QCX) hurt QRP Labs’ ability to fund future R&D. While the Pebble is a unique design, failing to mention the QCX hardware lineage can feel like “omission” to those who know the history.
Summary
If you are looking at this from a professional ethics lens: Yes, a mention of Hans Summers’ work on the QCX/QCX-mini PA and receiver architecture would be appropriate, as his specific implementation is the “DNA” that made the Pebble’s performance possible at that price.
8 Likes
Looks like fun. I’d be keen to try one side-by-side with my other rig. 
Now CW skill (yet) = muggle test (SSB).
Is it a trusdx or whatever in a different box, with a straight key bolted on?
I didn’t ask an AI. Thought I’d ask real people, which I presume you all are. 
5 Likes
Take the blue pill or the red one to find out!
2 Likes
We have so much AI forcefeeding at work I decided to retire.
We have an AI process that examines PDF files and compares against previous versions. Diffing PDF’s for changes is not terribly feasible because… so the AI differ is able to see real changes and list them. Ideal for technical documents where the format may change but the content is constant.
Last week a new PDF about a DRAM controller was “ingested” by the AI. It found there was a new signal “Yslerr” and generated a full description of this signal, what it meant, how the chip generated it and why. It was abrilliant looking summary. Except it was a hallucination. The new doc, a pre-release version, had a typing error. Somebody had mistyped “Ylserr” as “Yslerr” and the AI didn’t question if this was a typo but completely fabricated a description of a non-existent signal. Artifical stupidity perhaps.
Time to retire, wander around in pyjamas till 11am having a leisurely breakfast. And await the call if I fancy some consulting work helping fix AI errors 
6 Likes
That and the data lake isn’t being checked for validity, accuracy and currency by the HITL. Sounds like the model is ripe for over/underfitting.
Wait til 2027 for QAi and QaaS. Let the fun begin! Until then - Marbella, bed, pyjamas, leisurely breakfast and mark the occasional exam paper for a pension top up.
Happy retirement Andy whenever your last day is. 
1 Like
His last day might get kicked down the road a bit. He retired before, then discovered you can’t “re-tire” until you tire first …. and he wasn’t tired. Plus they offered a raise for him to come back part time. They’ll do it again, but label it consulting instead of part time.
Elliott, K6EL
1 Like
That takes me back. In the early ‘90’s I was working on Memory Test boards for 4M bit DRAM’s for a big (three letter) Computer manufacturer and the programming and sequencing of the signals was sometimes a bit complicated. Especially when we had to verify the different modes like Static Column and Fast Page mode with Read-Modify-Write thrown in as well. To say nothing of the address scrambling required to make them all physical “one’s” or “zero’s” for stressing the substrate over temperature. I’ve forgotten most of it now. I did get to see a 12” wafer though when I visited the customer site for testing, it was like a big shiny LP.
Andy
MM7MOX
My first thought was BBC, but I suspect that probably wasn’t the case… 
1 Like
Another thread that’s been diverted from the topic 
For those who are interested.
2 Likes
Why should I buy this which leverages the QCX with no real credit when I can buy a QCX and put cash into QRP Labs so they will design more original and innovative radios?
10 Likes
Andy, that’s your right. I only included this information for someone who wants to start in HF , not a seasoned pro like you.
It’s true, there’s also the Pixie. 
1 Like
Hi Kevin
uSDX was originally called QCX-SSB and was a very clever modification by Guido PE1NNZ to the original QCX (produced 2017-2020, replaced by the QCX+ and QCX-mini in 2020). Mostly Guido deleted large parts of the circuit and wrote new firmware for the ATmega328 processor. There are considerable performance penalties involved in uSDX: it’s not easy to design an SDR on an 8-bit 20MHz ATmega328 with 10-bit ADC. Later, many others were involved and Manuel DL2MAN developed uSDX further, with Guido PE1NNZ. There were many Chinese productions (I won’t call clones, since they did not breach any intellectual copyright, Manuel and Guido had not stipulated any license conditions in the beginning). Later Guido and Manuel tried to get some control back by producing (tr)uSDX which had a more guaranteed quality.
I believe Pebble is just a cut-down uSDX, making it single band removes some components. That and the fact they probably aren’t attempting to make a profit is why it ended up cheaper than uSDX.
The AI’s attempt isn’t bad! But “Hans spent years optimizing the specific component values (like the BS170 bias and the feedback loops) to make a $50 radio perform like a $500 one."… well the BS170 PA in the QCX doesn’t have any bias, nor any feedback And I certainly didn’t spend years developing it!
And “There has been historical friction because “clones” (unauthorized 1:1 copies of the QCX) hurt QRP Labs’ ability to fund future R&D.” - actually there has NEVER been any clone of the the QCX or QCX+ and QCX-mini, and my ability to fund future R&D had never been hurt in any way, let alone that way. So I’m not sure where AI made this up from! I haven’t had any friction with uSDX either. Actually when I designed QCX-mini, I had a spare bit of PCB area and I used it to make a little daughtercard which people wishing to modify it to a uSDX could populate with components and install in the QCX-mini. That and several 0-ohm resistors which would be removed in the modification, are STILL to this day part of the QCX-mini.
73 Hans G0UPL
12 Likes
Hi Hans 
, @G0UPL
Thanks for the clarification, Hans. In any case, I wouldn’t want to overshadow you, being a customer of yours I have several QRP Labs products made by you, and I just wanted to share information about this transceiver for those who might be interested in a small Pebble transceiver. (You’re well-listed on their website).
I haven’t yet tested the QMX (60m-15m) that I just received tax-free.
Truly
73, Éric F5JKK
(without AI 
)
2 Likes